Business Intelligence

Data Warehousing

The concept of data warehousing has existed for more than 30 years, however the development of methodologies and representation strategies developed within the last 30 or so years. Leaders in this area are Ralph Kimball and Bill Inmon who both defined methodologies and provided the representation substrate that defines a data warehouse.

A database designed to meet the requirements of a data warehouse provides a quick means of accessing large quantities of data and allows the developer to turn raw or aggregated data into usable information. The data is represented in a fashion that allows drilling down and drilling across the data to extract the information contained in the data.

At this time, the typical data warehouse contains tables with rows and columns to hold the data. This fits the typical model we have for data: a row represents a data point (a product purchase for instance) and its attributes (demographic data for instance). A collection of rows defined by attribute values can reveal a pattern of behavior ("women shopping for shoes are likely to buy double shot lattes at the coffee shop").

Knowledge Representation

The epistemological model of knowledge defines three areas:

• Knowledge by description or acquaintance - experience based knowledge
• Propositional Knowledge or "knowledge that"
• The knowledge of how - how to add to numbers together or how to drive to Oklahoma City.